Ab Initio Study of Photochemical Reactions of Ammonia Dimer

May 8, 1997 - Nobuhiro Okai, Akihiro Takahata, Masayuki Morita, Shinji Nonose, and ... Yoshiyuki Matsuda , Ken-ichiro Suhara , Naohiko Mikami , Asuka ...
0 downloads 0 Views 179KB Size
Download PDF
J. Phys. Chem. A 1997, 101, 3613-3618

3613

Ab Initio Study of Photochemical Reactions of Ammonia Dimer Systems Jong Keun Park* Department of Chemistry, Pusan National UniVersity, Pusan 609-735, Korea

Suehiro Iwata Institute for Molecular Science, Myodaiji, Okazaki 444, Japan ReceiVed: December 6, 1996; In Final Form: March 4, 1997X

Potential energy surfaces of ammonia dimer systems corresponding to the proton (hydrogen) transfer reaction from the reactant (NH3 + NH3) to the product (NH4 + NH2) have been calculated with ab initio HF, MP2, and SECI methods using a TZP basis set including diffuse Rydberg basis functions. The reaction path surfaces of the neutral, excited, and cationic states each have one local minimum corresponding to complexes such as (NH3‚‚‚NH3) for a neutral one, (H3N*-H‚‚‚NH2) for an excited one, and (H3N+-H‚‚‚NH2) for a cation, respectively. Dimer complexes such as (H3N-H‚‚‚NH2) and (H3N‚‚‚H-N+H2) do not exist. The photoionization and -dissociation reactions take place along the potential energy surfaces of the neutral, excited, and cationic states. The binding energies of the complexes relative to the asymptote are small. Particularly, the binding energies of H3N*-H‚‚‚NH2 (1A′′) and (H3N‚‚‚NH3) are only 0.24 and 0.16 eV, respectively. And the head-to-head type [(H3N‚‚‚NH3)+] with a binding energy of 1.52 eV does not exist on the protontransfer reaction surfaces.

1. Introduction Recently, experiments involving protonated and unprotonated ammonia clusters have been extensively performed.1-16 The photoionization mechanisms of ammonia clusters have been investigated by the various experimental techniques, such as single photon1-4 and multiphoton5-13 ionization and electron impact methods.14-16 The protonated cluster ions [(NH3)nH+] are found to be produced by two main steps.7,11,12 In the first one, the excited ammonia clusters (NH3)n-2(H3N*-H‚‚‚NH2) are formed from the predissociative state of the ammonia molecule in the cluster through the multiphoton absorption process. In the second one, the species (NH3)n-2 (H3N*-H‚‚‚NH2) releases an electron to produce the cluster ions, and then it is divided into the protonated ion and NH2 through a protontransfer reaction, while the unprotonated ion [(NH3)n+] is produced via absorption of a single photon or electron impact and subsequent release of an electron.4,14 Potential energy curves of protonated ammonia cluster ions via the multiphoton ionization mechanism have been constructed by Castleman et al.12,13 According to these potential energy curves, an electron is released by the multiphoton ionization process and the unprotonated cluster ions [(NH3)n+] dissociate into (NH3)n-1H+ + NH2 along the cationic potential curves. The (NH3)n+ and [(NH3)n-1H+ + NH2] species have local minima with an energy barrier. Very recently, these authors also measured experimentally the peaks of unprotonated cluster ions [(NH3)n+, n ) 2-5]. The peaks of these unprotonated ions are very weak compared to the protonated ions. In another experiment, Fuke et al.17 have suggested that protonated ammonia clusters can also be produced as a new photolysis product in the form of (NH2‚‚‚NH4)*(NH3)n-2. The intermediate (NH2‚‚‚NH4)*(NH3)n-2 is formed through a cage dynamics of the predissociation of large clusters. In this experiment, the protonated ion (NH2‚‚‚NH4+)(NH3)n-2 has been observed by single photon ionization of an intermediate having an ionization threshold of 3.79 eV. X

Abstract published in AdVance ACS Abstracts, April 1, 1997.

S1089-5639(96)04000-5 CCC: $14.00

It is worth noting that the theoretically adiabatic ionization energies for the unprotonated ammonia dimer cations are 8.518 and 7.71 eV,19 while the vertical energy is 9.15 eV.19 The experimental vertical energies are 9.541 and 9.19 eV.4 Although the photoionization process of the ammonia clusters has already been studied by various methods, further investigations seem to be worth carrying out on the basis of the following points. (i) In the dimer cation, does the head-to-head cationic complex (H3N‚‚‚NH3)+ dissociate into (NH4+ + NH2) along the proton-transfer reaction surface? That is, does the intermediate complex exist? (ii) Does a cage species (NH2‚‚‚NH4)* obtained as an intermediate of the intracluster reactions exist? (iii) Why were the unprotonated ammonia cluster ions [(NH3)n+] observed to be very weak except for n ) 1 and 2? (iv) In the ammonia dimer cation, what is the reason for the ionization energy gap between the theoretical and experimental values? Answers to these questions concerning a detailed understanding of the photoionization processes clearly require state-to-state correlation surfaces of the ground, excited, and cationic states between the reactant species (NH3 + NH3) and the products (NH4 + NH2). In this work, we have studied state-to-state correlation curves of the ground, excited, and cationic states from the reactant asymptote (NH3 + NH3) to product (NH4 + NH2) in order to obtain the relative energy gap and local intermediate complexes along the proton (hydrogen) transfer reaction paths. The protonated ammonia dimer cation is found to be produced from the excited complex of the hydrogen-bonded type by the absorption ionization dissociation [AID] or absorption dissociation ionization [ADI] mechanisms. Since studies of the mechanisms involving the unprotonated dimer cation seem to be quite limited theoretically and experimentally, we have also examined the ionization process of the nonhydrogen-bonded dimer cation (H3N‚‚‚NH3)+, which has a head-to-head type interaction. We show that the head-to-head interaction dimer cation plays an important role in the photoionization process of the unprotonated ammonia dimer cation. In addition, the photodissociation reactions of the NH4 radical are examined as a fundamental © 1997 American Chemical Society

3614 J. Phys. Chem. A, Vol. 101, No. 19, 1997

Figure 1. Potential energy curves for the ground and excited states of the NH4 Rydberg radical dissociating into (NH3 + H).

unit in photochemical processes of (NH3)n clusters. Our results are also compared with the experimental20-22 and other theoretical23-33 values. 2. Calculational Methods The basis sets chosen are the triple-ζ basis on N (521/2111)45 and H(511).46 Two extra d type polarization functions45 are added to nitrogen (Rd ) 0.412, 1.986) and one extra p type function to hydrogen (Rp ) 0.990 495). The diffuse Rydberg basis functions (Rs ) 0.028, 0.0066; Rp ) 0.025, 0.0051; Rd ) 0.015, 0.0032)47 are additionally augmented on nitrogen to describe the Rydberg states of NH3 and NH4. The total number of contracted basis set used in ammonia dimer is 130. The geometric structures of the ammonia monomer, dimers, and their cations are optimized with the MP2 (second-order Mo¨ller Plesset) approximation using Gaussian 92. The potential energy surfaces of the proton-transfer reaction from (NH3 + NH3+) to (NH4+ + NH2) via a complex (H3N+-H‚‚‚NH2) is calculated with the UHF wave functions for RNN and rNH by optimizing the other geometric parameters. The excited states of NH3, (NH3)2, and NH4 are somewhat of a Rydberg nature with a cationic core. The ground state of NH4 itself is often called the Rydberg radical. Therefore, the geometric structures of these states are expected to be similar to those of the corresponding cations. To draw the potential energy curves and surfaces, we use the characteristics of the states twofold. The geometry of the corresponding cation is assumed in drawing the potential energy curves for the dissociation of the ground and excited states of NH4 to (NH3 + H) and for the hydrogen transfer reactions of (NH3 + NH3*) to (NH4* + NH2) via a complex (H3N*-H‚‚‚NH2). Besides, the molecular orbitals for the configuration interaction (CI) are determined with the open shell restricted Hartree-Fock method for the corresponding cation. The SECI (singly excitation configuration interaction) method is used for the potential energy calculations. To examine the appropriateness of the procedure, the potential energies of the ground-state surface from NH4 to (NH3 + H) have been calculated with the SDCI and CCSD(T) methods. In the SDCI and CCSD(T) calculations an f type basis function (Rf ) 1.093)48 is added to the above basis set. 3. Results and Discussion The potential energy curves for the ground and low lying excited states of the NH4 Rydberg radical dissociating into (NH3 + H) are drawn in Figure 1. Because of the complexity, we

Park and Iwata have cut the potential energy curves of the high-lying excited states at RNH )1.2 Å and have not connected the curves between RNH ) 4.0 Å and the asymptote (NH3 + H). The 2A1 ground state of the NH4 radical correlates with the asymptote (NH3 + H). The ground state is metastable, and the potential well is shallow. Although the ground state of the NH4 radical has an energy barrier of 0.6 eV along the NH bond rupture, the stability of the NH4 radical seems to be influenced by tunneling. The existence of the NH4 radical has been confirmed by the solvated electron reaction and by electrochemical approaches.34-37 The lifetime of the NH4 radical was measured experimentally to be 13 ps, a value more than 106 times shorter than in the one for NH4 in ammonia clusters.17 The equilibrium and transition state distances and the relative energies of the NH4 radical dissociating into (NH3 + H) are listed in Table 1 together with the ionization and excitation energies of the ammonia molecules. All energies are adiabatic values. The excitation energies of the ammonia molecules [NHn (n ) 2-4)] calculated by the SECI method are in reasonable agreement with the experimental20-23 and theoretical24,26,27 values. The NH equilibrium internuclear distance of ∼1.04 Å is larger than that (RNH ) 1.012 Å) of a monomer NH3, since in its ground state the NH4 radical has an electron in a 3s Rydberg orbital, that is, NH4 is a semi-ionic state. At the transition state the bond length (RNH) is ∼1.43 Å, that is, the bond breaking in NH4 takes place near the equilibrium geometry. Therefore, the NH4 radical has the weak NH bonds. The energy barriers of the NH4 and (NH3 + H) to the transition state are ∼0.8 and ∼0.6 eV, respectively. In the excitation energy of the NH4 radical, our results are similer to the other theoretical24,26,27 and experimental21,22 results. State-to-state correlation diagrams of the ground, excited, and cationic state surfaces correlating the dimer reactant asymptote (NH3 + NH3) and the product asymptote (NH4 + NH2) are drawn in Figure 2. All energy gaps are adiabatic. The potential energy of the ammonia dimer is set equal to zero. The complexes and the asymptotes are connected with dotted lines. The ammonia dimer complexes on these surfaces correspond to NH3‚‚‚NH3, H3N*-H‚‚‚NH2, and H3N+-H‚‚‚NH2 for the neutral, the excited, and the cationic complexes, respectively. Similar to previous results,38-44 we also found that the structure of the dimer is not a hydrogen-bonded type complex. The floppy nature of this complex can be associated with three motions, that is, internal rotation of each monomer, interchange tunneling, and inversion tunneling of each monomer. The N‚‚‚H internuclear distance of the dimer calculated with the MP2 method is 2.306 Å. The RNN distance in this work and the experimental40 one are 3.326 and 3.337 Å, respectively. The binding energies of the ammonia dimer complex (NH3‚‚‚NH3) from the asymptotes (NH3 + NH3) and (NH4 + NH2) are 0.16 and 4.87 eV, respectively. From the complex (NH3‚‚‚NH3) to the asymptote (NH4 + NH2), the hydrogen transfer reaction takes place through the ground-state surface. We could not find the local minimum of a complex such as (H3N-H‚‚‚NH2) on the surface. On the first excited-state surface correlating [NH3 + NH3*˜ 2A1)], our local minimum corresponds (1A1)] to [NH4 + NH2*(A to an excited complex (H3N-H‚‚‚NH2)*(1A′). From the complex (H3N-H‚‚‚NH2)* to the asymptotes (NH3 + NH3*) and (NH4 + NH2*), the hydrogen transfer reaction also takes place through the excited-state surface. The binding energies of this ˜ 2A1)] and excited complex from the asymptote [NH4 + NH2*(A 1 [NH3 + NH3*( A1)] are 0.75 and 0.95 eV, respectively. The

Ammonia Dimer Systems

J. Phys. Chem. A, Vol. 101, No. 19, 1997 3615

TABLE 1: Relative Energies (eV) for the 2A1 Ground States along the NH4 Radical Dissociating into NH3 + Ho R(NH)eqg R(NH)TSg ∆E[TS-(NH3+H)] ∆E[NH4-(NH3+H)] ∆E(NH4-TS) NH4 IEi ∆E(3s-3p) ∆E(3s-4s) ∆E(3s-3d); 2T2 ∆E(3s-3d); 2E ∆E(3s-4p) ∆E(3s-5s) ∆E(3s-4d); 2T2 ∆E(3s-4d); 2E ∆E(3p-3d); 2T2 NH3 IEj ∆E(n-3s); A ˜ 3A1 ∆E(n-3px,y); B ˜ 3E ∆E(n-3pz); 3A1 ∆E(n-4s); 3A1 ∆E(n-3d); 3E NH2 ∆E(n-3s); A ˜ 2A1

SECIa

QDCIb

CCSD(T)c

CIPSId

SDCIe

1.022 1.339 0.59 -0.17 0.75

1.040 1.439 0.61 -0.22 0.83

1.040 1.425 0.57 -0.22 0.79

1.033 1.369 0.52 -0.23 0.88

1.041 1.427 0.64 -0.21 0.85

1.55 2.46 2.61 2.69 2.90 3.22 3.29 3.32 1.06

1.90 2.78 2.99 3.13 3.31

1.66 2.65

1.89

6.46 7.88 8.29 8.98 9.09

6.31 7.86 8.05 9.06

1.40

1.29

exptlf

-0.3h

4.62k

4.60 2.89 3.04

2.18l

1.87l

1.23 10.14 6.27m 7.84m 7.84m

10.17 6.38 7.90 8.14 9.11 1.26n

a SECI energies were obtained with the MOs and geometries of NH4+ calculated by RHF at each internuclear distance. b QDCI energies were obtained with a suite of MOLYX. c CCSD(T) energies were obtained with Gaussian 92. d Reference 27. e Reference 26. f Reference 20. g Unit of internuclear distance is angstrom. h Cited from ref 27. i Ionization energy of NH4 is optimized with MP2 calculations. j Ionization energy of NH3 is optimized with MP2 calculations. k Reference 21. l Reference 22. m Reference 24. n Reference 23. o All energies are adiabatic values.

Figure 2. Schematic energy diagram of the hydrogen and proton transfer surfaces from the asymptote (NH3 + NH3) to (NH4 + NH2), and relative energies of the neutral, excited, and cationic complexes. All energies are adiabatic values and are in units of eV. Superscripted letters indicate data from the following sources: (a) ref 1; (b) ref 4; (c) ref 20; (d) ref 16; (e) ref 18; (f) ref 17; (g) ref 23; (h) ref 28.

energy gaps of (H3N-H‚‚‚NH2)* relative to (H3N+-H‚‚‚NH2; 2A′′) and (NH ‚‚‚NH ) are ∼2.93 and ∼5.52 eV, respectively. 3 3 We note that Fuke et al.17 have found a cage product in the predissociation process of large ammonia clusters and assigned it as an excited-state (NH4*‚‚‚NH2)(NH3)n-2. An ionization threshold of 3.79 eV by a single photon mechanism from this complex has been observed. This ionization threshold is larger than our energy gaps of 2.93 eV. Therefore, the cage complex is different from the first excited-state (H3N-H‚‚‚NH2)*(1A′). The potential energy surface of the reaction [(NH3 + NH3*; 1A ), A] f [(H N-H‚‚‚NH )*(1A′), B] f [(NH + NH ; A 2 1 3 2 4 2 ˜ A1), C] is presented in Figures 3 and 4. The contour map drawn corresponds to potential energies for internuclear distances ranging from 2.4 to 3.4 Å for NN and from 0.7 to 2.7 Å for NH. To understand easily the reaction path surface, we have just represented A and C on Figures 3 and 4. A and C are not the positions of the asymptotes. A circle near A is an artifact

caused by the fitting procedure of the potential energy curves. This contour map has a local minimum point at B. The distances of RNN and rNH at B are 3.09 and 1.07 Å, respectively. Figure 4 is drawn by cutting the vertical axis of Figure 3. The second excited complex (H3N*-H‚‚‚NH2; 1A′′) correlates with [NH4*(2T2) + NH2(X2B1)]. From the complex (H3N*H‚‚‚NH2) to the asymptotes (NH3 + NH3*) and (NH4* + NH2), the proton and charge transfer reactions take place through the excited-state surface. This excited complex had been found to be formed by a predissociative reaction (NH3 f NH2 + H) of the ammonia monomer in the ammonia clusters via an intracluster reaction. The binding energy of this state relative to the asymptote [NH4*(2T2) + NH2(X2B1)] is 0.24 eV. The potential well is very shallow. Thus, the complex (H3N*-H‚‚‚NH2) is easily dissociated into NH4* + NH2. The energy gaps of (H3N*-H‚‚‚NH2; 1A′′) relative to (H3N+-H‚‚‚NH2; 2A′′) and (NH3‚‚‚NH3) are ∼1.92 and ∼6.53 eV, respectively.

3616 J. Phys. Chem. A, Vol. 101, No. 19, 1997

Figure 3. Energy contour map of the hydrogen transfer surface covering the reactant [NH3 + NH3*(1A1), A], the complex [(H3NH‚‚‚˜ 2A1), C]. The energy NH2)*(1A′), B], and the asymptote [NH4 + NH2(A difference between the neighboring isoenergetic lines is 0.112 eV. A and C are not the exact positions of the asymptotes.

Figure 4. Reaction path surface of the hydrogen transfer reaction covering the reactant [NH3 + NH3*(1A1), A], the complex [(H3NH‚‚‚˜ 2A1), C]. NH2)*(1A′), B], and the asymptote [NH4 + NH2 (A

Figure 5. Energy contour map of the proton (charge) transfer surface covering the reactant [NH3 + NH3*(1E), A], the complex [(H3N*H‚‚‚NH2; 1A′′), B], and the asymptote [NH4*(2T2) + NH2(X2B1), C]. The energy difference between the neighboring isoenergetic lines is 0.1299 eV. A and C are not the exact positions of the asymptotes.

We have also calculated the second excited potential energy surface of the reaction [NH3 + NH3*(1 E), A] f [(H3N*-H‚‚‚NH2; 1A′′), B] f [NH4* + NH2(X2B1), C], which is shown in Figures 5 and 6. The contour map drawn corresponds to potential energies for internuclear distances ranging from 2.4 to 3.4 Å for NN and from 0.7 to 2.7 Å for NH. To understand easily the reaction path surface, we have just represented A and C on Figures 5 and 6. A and C are not the positions of the asymptotes. B is a minimum complex (H3N*-H‚‚‚NH2; 1A′′) at this surface. The potential energy increases directly from B to A and C. The RNN and rNH at B are 2.93 and 1.03 Å, respectively. Figure 6 is drawn by cutting the vertical axis of Figure 5. The energy gap between B and C is small, and this excited complex is slightly unstable. Therefore, in the process of producing the protonated ammonia cluster cation, the competing process between the photon absorption-ionizationdissociation (AID) and the photon absorption-dissociation-

Park and Iwata

Figure 6. Reaction path surface of the proton (charge) transfer reaction covering the reactant [NH3 + NH3*(1E), A], the complex [(H3N*H‚‚‚NH2; 1A′′), B], and the asymptote [NH4*(2T2) + NH2(X2B1), C].

ionization (ADI) mechanism in two-photon ionization will take place. This state plays a bridge role in the resonance-enhanced multiphoton ionization mechanism. There are two possible geometric structures for the dimer cation. One is a hydrogen-bonded dimer cation (H3N+-H‚‚‚NH2; 2A′′) with binding energies of 0.97 and 1.85 eV relative to the [NH4+(1A1) + NH2] and [NH3 + NH3+(2A2′′)] asymptotes, respectively. The energy gap between this dimer cation (2A′′) and the second excited complex (H3N*-H‚‚‚NH2; 1A′′) is 1.92 eV, and the ionization energy of this complex relative to the neutral dimer (NH3‚‚‚NH3) is 8.45 eV. Ionization energy of 8.5 eV for the unprotonated ammonia dimer cations suggested by Kassab et al.18 is similar to ours. Therefore, they might have calculated the ionization energy of the hydrogen-bonded type dimer cation (H3N+-H‚‚‚NH2; 2A′′). The other is a head-to-head interaction (H3N‚‚‚NH3)+ (D3d symmetry), which is formed by the interaction between the dipole moment and the cationic charge. According to the geometric structure of this complex, the ionization of the NH3 occurs with the single photon, and NH3+ combines with NH3. The ionization energy of this cation relative to (NH3‚‚‚NH3) is 8.78 eV. The binding energy of (H3N‚‚‚NH3)+ with respect to (NH3 + NH3+) is 1.52 eV. The potential energy of (H3N‚‚‚NH3)+ is higher than that of (H3N+-H‚‚‚NH2). The cationic potential energy surfaces for the proton-transfer reaction [NH3 + NH3+(2A2′′), A] f [H3N+-H‚‚‚NH2; 2A′′), B] f [NH4+(1A1) + NH2(X2B1), C] are drawn in Figures 7 and 8. The contour map drawn corresponds to potential energies for internuclear distances ranging from 2.4 to 3.6 Å for NN and from 0.7 to 2.9 Å for NH. To understand easily the reaction path surface, we have just represented A and C on Figures 7 and 8. A and C are not the positions of the asymptotes. B is only a local minimum point in this potential energy surface. The optimized distances RNN and rNH at B are 2.87 and 1.04 Å, respectively. A small circle near A is an artifact caused by the fitting procedure. The potential curve increases stiffly from B to A. The binding energy between B and C is slightly larger than those of the excited states. The contour map of the dimer cation is very similar to that of the excited molecule. Figure 8 was drawn by cutting the vertical axis of Figure 7. In Figure 8, B is also represented as a minimum point. Because Figure 8 was drawn with potential energies for RNN ranging from 2.6 to 3.6 Å, the energy gaps of B relative to A and C are different from that of Figure 7. We note that Castleman et al.13 observed very weak peaks of the unprotonated ammonia cluster [(NH3)n+ (n ) 2-5)] using the multiphoton ionization method13 and drew a schematic diagram of the potential energy curves during multiphoton ionization of the protonated ammonia clusters.12 The peaks of

Ammonia Dimer Systems

Figure 7. Energy contour map of the proton transfer surface covering the reactant [NH3 + NH3+(2A2′′), A], the complex [(H3N+H‚‚‚NH2; 2 A′′), B], and the asymptote [NH4+(1A1) + NH2(X2B1), C]. The energy difference between the neighboring isoenergetic lines is 0.10 eV. A and C are not the exact positions of the asymptotes.

J. Phys. Chem. A, Vol. 101, No. 19, 1997 3617 the second excited state. In the cationic energy curve the proton transfer takes place. According to the geometric structure, the head-to-head interaction dimer cation (H3N‚‚‚NH3)+ does not exist on the proton-transfer reaction surfaces from the reactant (NH3 + NH3+) to the product (NH4+ + NH2). But the protonated dimer cation of (H3N+-H‚‚‚NH2; 2A′′) exists on the proton transfer reaction surface. The energy barrier between the two isomers is small. Therefore, the geometric change from the head-tohead to the hydrogen-bonded type takes place easily with an excess energy. The ionization energies of (H3N‚‚‚NH3)+ and (H3N+-H‚‚‚NH2; 2A′′) relative to the neutral ammonia dimer are 8.78 and 8.45 eV, respectively. The first excited state (H3N-H‚‚‚NH2)*(1A′) correlates with ˜ 2A1)] including the binding energy of 0.75 eV. [NH4 + NH2(A The second excited state (H3N*-H‚‚‚NH2; 1A′′) correlates with [NH4*(2T2) + NH2] by its binding energy of 0.24 eV. Because of the weak binding, a competing reaction between AID and ADI takes place to produce the protonated ammonia dimer cation. These excited states play an important bridge role on the multiphoton ionization process. Acknowledgment. The authors thank Professor Fuke and Dr. Ten-no for their valuable discussions. The Korea Science and Engineering Foundation is acknowledged for the financial support of Jong Keun Park. References and Notes

Figure 8. Reaction path surface of the proton transfer reaction covering the reactant [NH3 + NH3+(2A2′′), A], the complex [(H3N+H‚‚‚NH2; 2A′′), B], and the asymptote [NH +(1A ) + NH (X2B ), C]. 4 1 2 1

the unprotonated ammonia clusters were very weak except for the dimer cation. In our results, a head-to-head type dimer cation (H3N‚‚‚NH3)+ does not exist on the proton transfer reaction surfaces from the reactant to the product. The protonated and unprotonated clusters have two local minima by an energy barrier of 0.45 eV. The energy gap between the two isomers is 0.33 eV. The head-to-head type structure might contribute to the peak of the unprotonated dimer cation with a binding energy of ∼1.52 eV. The geometric structure of the head-to-head is different from the protonated dimer cation (hydrogen-bonded). To change the geometric structure from the head-to-head to the hydrogen-bonded type, a monomer is rotated to the other, and a proton of the monomer transfer to the other. Because of the small energy barrier, the head-tohead complex may be transferred to the hydrogen-bonded type if the excess energy is supplied by either electron or photon excitation. 4. Conclusions We have calculated state-to-state correlation energy curves for the hydrogen and proton transfer reactions from the reactant asymptote (NH3 + NH3) to the product (NH4 + NH2). The correlation energy curves of the neutral, excited, and cationic species have only one local minimum complex each. Each local minimum on the potential surfaces corresponds to complexes such as (H3N+-H‚‚‚NH2) for the cation, (H3N*-H‚‚‚NH2) for the excited state, and (NH3‚‚‚NH3) for the neutral. But the dimer complexes such as (H3N-H‚‚‚NH2), (H3N‚‚‚H-N+H2), and (H3N‚‚‚H-N*H2) have not been confirmed. On the energy curves of the neutral and first excited states the hydrogen (atom) transfer reaction takes place. But the proton and charge transfers take place on the energy curve of

(1) Ceyer, S. T.; Tiedemann, P. W.; Mahan, B. H.; Lee, Y. T. J. Chem. Phys. 1979, 70, 14. (2) Shinohara, H.; Nishi, N.; Washida, N. J. Chem. Phys. 1985, 83, 1939. (3) Kaiser, E.; de Vries, J.; Steger, H.; Menzel, C.; Kamke, W.; Hertel, I. V. Z. Phys. D 1991, 20, 193. (4) Kamke, W.; Herrman, R.; Wang, Z.; Hertel, I. V. Z. Phys. D 1988, 10, 491. (5) Shinohara, H.; Nishi, N. Chem. Phys. Lett. 1982, 87, 561. (6) Shinohara, H. J. Chem. Phys. 1983, 79, 1732. (7) Echt, O.; Dao, P. D.; Morgan, S.; Castleman, A. W., Jr. J. Chem. Phys. 1985, 82, 4076. (8) Wei, S.; Tzeng, W. B.; Castleman, A. W., Jr. J. Chem. Phys. 1990, 92, 332. (9) Pursell, C. J.; Weliky, D. P.; Oka, T. J. Chem. Phys. 1990, 93, 7041. (10) Wei, S.; Tzeng, W. B.; Castleman, A. W., Jr. J. Chem. Phys. 1990, 93, 2506. (11) Misaizu, F.; Houston, P. L.; Nishi, N.; Shinohara, H.; Kondow, T.; Kinoshita, M. J. Chem. Phys. 1993, 98, 336. (12) Purnell, J.; Wei, S.; Buzza, S. A.; Castleman, A. W., Jr. J. Phys. Chem. 1993, 97, 12530. (13) Buzza, S. A.; Wei, S.; Purnell, J.; Castleman, A. W., Jr. J. Chem. Phys. 1995, 102, 4832. Wei, S.; Purnell, J.; Buzza, S. A.; Castleman, A. W., Jr. J. Chem. Phys. 1993, 99, 755. (14) Odutola, J. A.; Dyke, T. R.; Howard, B. J.; Muenter, J. S. J. Chem. Phys. 1979, 70, 4884. (15) Stephan, K.; Futrell, J. H.; Peterson, K. I.; Castleman, A. W., Jr.; Wager, H. E.; Djuric, N.; Mark, T. D. Int. J. Mass Spectrom. Ion Phys. 1982, 44, 167. (16) Posey, L. A.; Guettler, R. D.; Kirchner, N. J.; Zare, R. N. J. Chem. Phys. 1994, 101, 3772. (17) Fuke, K.; Takasu, R. Bull. Chem. Soc. Jpn. 1995, 68, 3309. (18) Cao, H. Z.; Evleth, E. M.; Kassab, E J. Chem. Phys. 1984, 81, 1512. (19) Greer, J. C.; Ahlrichs, R.; Hertel, I. V. Z. Phys. D 1991, 18, 413. (20) Skerbele, A.; Lassettre, E. N. J. Chem. Phys. 1965, 42, 395. (21) Fuke, K.; Takasu, R.; Misaizu, F. Chem. Phys. Lett. 1994, 229, 597. (22) Herzberg, G. Faraday Discuss. Chem. Soc. 1981, 71 165. (23) Lathan, W. A.; Hehre, W. J.; Curtiss, L. A.; Pople, J. A. J. Am. Chem. Soc. 1971, 93, 6377. (24) Runau, R.; Peyerimhoff, S. D.; Buenker, R. J. J. Mol. Spectrosc. 1977, 68, 253. (25) Albritton, D. L. At. Data Nucl. Data Tables 1978, 22, 1. (26) Kaspar, J.; Smith, V. H., Jr.; McMaster, B. N. Chem. Phys. 1985, 96, 81. (27) Kassab, E.; Evleth, E. M. J. Am. Chem. Soc. 1987, 109, 1653.

3618 J. Phys. Chem. A, Vol. 101, No. 19, 1997 (28) Amor, N. B.; Maynau, D.; Spiegelmann, F. J. Chem. Phys. 1996, 104, 4049. (29) Tachibana, A.; Kawauchi, S.; Kurosaki, Y.; Yoshida, N.; Ogihara, T.; Yamabe, T. J. Phys. Chem. 1991, 95, 9647. (30) Gill, P. M. W.; Radom, L. J. Am. Chem. Soc. 1988, 110, 4931. (31) Tachikawa, H.; Tomoda, S. Chem. Phys. 1994, 182, 185. (32) Tomoda, S.; Kimura, K. Chem. Phys. Lett. 1985, 121, 159. (33) Bouma, W. J.; Radom, L. J. Am. Chem. Soc. 1985, 107, 345. (34) Wan, J. K. S. J. Chem. Educ. 1968, 45, 40. (35) Brooks, J. M.; Dewald, R. R. J. Chem. Phys. 1971, 75, 986. (36) Kariv-Miller, E.; Nanjundiah, C.; Eaton, J.; Swenson, K. E. J. Electroanal. Chem. 1984, 167, 141. (37) Evleth, E. M.; Kassab, E. Pure Appl. Chem. 1988, 60, 209. (38) Snels, M.; Fantoni, R.; Sanders, R.; Weerts, W. L. Chem. Phys. 1987, 115, 79. (39) Havenith, M.; Linnartz, H.; Zwart, E.; Kips, A.; ter Meulen, J. J.;

Park and Iwata Meerts, W. L. Chem. Phys. Lett. 1992, 193, 261. (40) Olthof, E. T. H.; van der Avoird, A.; Wormer, P. E. S. J. Mol. Struct. 1994, 307, 201. (41) Linnartz, H.; Meerts, W. L.; Havenith, M. Chem. Phys. 1995, 193, 327. (42) Sagarik, K. P.; Ahlrichs, R.; Brode, S. Mol. Phys. 1986, 57, 1247. (43) Greer, J. C.; Ahlrichs, R.; Hertel, I. V. J. Chem. Phys. 1991, 95, 3861. (44) Greer, J. C.; Ahlrichs, R.; Hertel, I. V. Z. Phys. D 1991, 18, 413. (45) Huzinaga, S.; Andzelm, J.; Klobukowski, M.; Radzio-Andzelm, E.; Sakai, Y.; Tatewaki, H. Physical Science Data, Gaussian basis sets for molecular calculations; Elsevier: Amsterdam, 1984; Vol. 16. (46) Hasimoto, K.; Osamura, Y. J. Chem. Phys. 1991, 95, 1121. (47) Poirier, R.; Kari, R.; Csizmadia, I. G. Physical Science Data, Handbook of Gaussian basis sets; Elsevier: Amsterdam, 1985; Vol. 24. (48) Almlo¨f, J.; Taylor, P. R. J. Chem. Phys. 1987, 86, 4070.